BF-2: Quantifying carbon storage
Some of the strategies for climate change mitigation are aimed to protect, restore and/or manage forested ecosystems that sequester and store high amounts of carbon (e.g., Kauppi et al. 2022, Nunes et al. 2019, Fargione et al. 2018, Armenteras et al. 2015, Lemprière et al. 2013), providing some directions for policy and opportunities in carbon markets (e.g., Anderson et al. 2017, but see Venter & Koh 2012, West et al. 2023). However, soil carbon storage (Spawn et al. 2020), an important global asset in northern ecosystems, in particular peatlands in Canada (Sothe et al. 2021, Harris et al. 2021), is crucial to maintain global climate stability (Chaplin-Kramer et al. 2022, Neugarten et al. 2023).
Measuring and modeling the past and current redistribution and storage of anthropogenic carbon sources among the atmosphere, ocean and terrestrial biosphere is key to understanding the global carbon cycle and informing global decision making (Friedlingstein, 2020; Global Carbon Project). In terrestrial ecosystems, peatlands also called “...organic soils”, “bogs”, “fens”, “swamps” and “mires”, are the world’s most carbon-dense terrestrial ecosystems.
Peatlands are formed from partially decomposed plant remains that have accumulated over thousands of years under conditions of waterlogging. Peat soils hold an estimated 650 billion tonnes (Gt = Pg) of carbon on only 3 percent of the Earth’s land surface – a carbon store that is equal in magnitude to the amount of carbon in the Earth’s vegetation, and more than half of the carbon in the atmosphere (Xu et al. 2010, Page et al. 2011, Dargie et al. 2017)….” FAO-UN (2020). The Hudson James Bay Lowlands (HJBL) is the second largest peatland in the world, which extends over ~370,000 km2 (Harris et al. 2021, Packalen et al. 2014, WCS story maps). Nichols & Peetet (2019) argued that the estimate of total northern peat carbon stocks could be from 545 Gt to 1,055 Gt of carbon.
“One-quarter of the world’s northern peatlands (~1.1 million km2), and the world’s largest peatland C stock (~150 Gt), is located in Canada (Joosten 2009, Xu et al. 2018, Hugelius et al. 2020)…only ~10% of peatlands in Canada are currently within protected areas” (Harris et al. 2021). Soil carbon is critically important in the northern hemisphere and in particular for Canada (Schalermann et al. 2014, Beaulne et al. 2021).
Spatial variability in carbon stored in peatlands complicates upscaling to regional estimates. For instance “regions characterized by warmer and wetter conditions stored the most C as peat” (Packalen et al. 2016). Combining field measurements, satellite data, hydrological and climate data is essential to get better spatially-explicit estimates and reduce uncertainty (seeSothe et al. 2022, Balogun et al. 2023). Similarly, implemeting soil carbon simulation models and identifying areas of improvement would reduce uncertainty in carbon dynamics under climate change (Varney et al. 2022).
Here we present three general approaches for quantifying carbon storage for ecosystem services assessments and show how these three approaches could be adapted to account for peatland carbon dynamics. All these approaches rely on local site-specific field surveys for calibration and validation.
Stratify and multiply
The approach relies on first stratifying or partitioning the study area into multiple unique land cover or ecosystem types (Goetz et al. 2009, Sharp et al. 2020, Martínez-López et al. 2019). A reference or average carbon stock value is assigned to each cover type. The area of each cover type is then multiplied by the cover type’s associated reference carbon stock value.
Many ecosystem service modeling platforms use a Stratify & Multiply approach, including the ARtificial Intelligence for Ecosystem Services, ARIES (Martínez-López et al. 2019) and the Integrated Valuation of Ecosystem Services and Tradeoffs, InVEST (Natural Capital Project, 2023). Best practices include using local site-specific field surveys to derive reference carbon stock values and detailed site-specific land cover maps that account for environmental gradients that might influence carbon stock changes (e.g., climate, soil characteristics, species composition, management, degradation, elevation, site history, wetland type, ecozone, etc.).
Carbon emissions can be assessed in a similar way. For example, maps of peatland types can be combined with estimates of Net Ecosystem Exchange and Methane emission rates compiled for each peatland type and region (Webster et al. 2018).
Assessment of Status
Some Global and Canadian examples of model implementation.
| Models/Approaches | Have they been used? | In Canada? |
|---|---|---|
| Stratify and multiply | Goetz et al. (2009) |
Direct remote sensing
The second approach, Direct Remote Sensing, uses a statistical or machine learning model to relate soil carbon field measurements to remotely sensed observations and other geospatial layers (Goetz et al. 2009, Sothe et al. 2021). The Direct Remote Sensing approach captures spatial variability of carbon within land cover classes, which is often missed with the Stratify and Multiply approach. Best practices include using local site-specific field surveys for model calibration and validation as well as local site-specific geospatial layers for covariates (e.g., topography, slope, soil type, wetland type, ecosystem type, soil characteristics, ecozone, etc.).
Carbon stock maps can be differentiated to track changes over time. Another new approach relies on multi-temporal Light Detection and Ranging (LiDAR) to estimate changes in peatland elevation after a wildfire (Reddy et al. 2015). Peatlands are becoming more susceptible to wildfires due to lowering water tables under climate change and development (Turetsky et al. 2015).
Assessment of Status
Some Global and Canadian examples of model implementation.
| Models/Approaches | Have they been used? | In Canada? |
|---|---|---|
| Direct remote sensing | Goetz et al. (2009) | Sothe et al. (2021) |
Mass balance models
The third approach, Mass balance models, accounts for all key carbon pools and flows of carbon between these pools over time (GFOI 2016, Bona et al. 2020). Mass balance models can track changes in carbon due to natural processes such as growth and decay as well as drivers of change including land use, wildfires, water table depth, and permafrost thaw.
Mass balance models are especially useful for understanding peatland carbon dynamics because they can account for the balance between the uptake of carbon and the emission of methane (Bona et al. 2020). Furthermore, these models can track changes in carbon due to increases in peatland disturbances, including droughts, permafrost thaw, and fires under climate change (Bona et al. 2020, Harris et al. 2022). When peatland permafrost thaws, previously frozen carbon can be decomposed, thereby releasing carbon dioxide and methane (Hugelius et al. 2020).
Many modeling approaches have been implemented to estimate peatland carbon stocks in Canada including (1) the Canadian model for peatlands (CaMP) which accounts for wildfire disturbances but currently not permafrost thaw (Bona et al. 2020, although permafrost thaw is currently being implemented), (2) dynamic global ecosystem models that incorporate permafrost dynamics (Chaudhary et al. 2017), and (3) inventory-based models (Hugelius et al. 2020) that can also account for abrupt permafrost thaw (Turetsky et al. 2020). Best practices include integrating local site-specific field data for calibration and validation as well as remotely sensed data for accurate spatially explicit disturbance detection (e.g., Zhou et al. 2021, Bona et al. 2020, Shaw et al. 2021).
Assessment of Status
Some Global and Canadian examples of model implementation.
| Models/Approaches | Have they been used? | In Canada? |
|---|---|---|
| Mass balance Models | GFOI, (2016) |
Additional models to evaluate vulnerabilities and monitoring peatlands
Projections, vulnerabilities
Assessment of Status
Some Global and Canadian examples of model implementation.
| Models/Approaches | Have they been used? | In Canada? |
|---|---|---|
Projections/ vulnerabilities |
Müller & Joos (2021), |
Mapping & monitoring
Assessment of Status
Some Global and Canadian examples of model implementation.
| Models/Approaches | Have they been used? | In Canada? |
|---|---|---|
| Mapping and monitoring |
No identify yet
| People/Organization/Institution | Topic |
|---|---|
| University of Waterloo (CanPeat) | Canadian Model for Peatlands (CaMP) Can-Peat: Canada’s peatlands as nature-based climate solutions |
| ECCC (Kelly Bona) | |
| WWF | Mapping |
| WCS | Mapping |
| University of Toronto (Sarah Finkelstein) | |
| University of Leads | PEATMAP |
| Stockholm University (Bolin Centre for Climate Research) |
In construction!!!
| Description | Repository and Layers | Extent, Format, Resolution, Projection |
Reference |
|---|---|---|---|
| Maps of northern peatland extent, depth, carbon storage and nitrogen storage | Stockholm University (Bolin Centre for Cliamte Research) Histel_fraction Histel_minerotrophic Histel_ombrotrophic Histel_SOC_hg_per_sqm Histel_TN_gram_per_sqm Histosol_fraction Histosol_minerotrophic Histosol_ombrotrophic Histosol_SOC_hg_per_sqm Histosol_TN_gram_per_sqm Mean_potential_peat_depth_cm |
Norhtern hemisphere, GeoTiff, 10 Km, World Azimuthal Equidistant projection, |
Hugelius et al. (2020) |
| A map of global peatland extent created using machine learning (Peat-ML) |
Peatland fractional coverage |
Global, NetCDF (Network Common Data Form), 0.0833 (~1Km), lonlat |
Melton et al. (2022) |
| PEATMAP: Refining estimates of global peatland distribution based on a meta-analysis | Global (by continents) Shapefile 25m-1000m Projected (ESRI:54034 - World_Cylindrical_Equal_Area) |
Xu et al. (2018) | |
| Global Peatlands |
Combine various data sets to delineated peatlands Available in tiles |
Global TIF 30 * 30m EPSG:4326 - WGS 84 |
Global Forest Watch |
Example
Peat ML global peatland_extent from Melton et al. (2022)
Peat cores Hugelius et al. (2020)
There is a need to optimize monitoring/sampling in the region by identifying priority sites where data is needed to improve model performance and reduce uncertainty.
Partnerships with local NGOs, universities/researchers, and indigenous communities are needed to support the implementation of an observing system.